A regional soil pollution health risk assessment method based on POI data

Abstract

This invention relates to the field of regional soil pollution health risk assessment technology, and more particularly to a method for regional soil pollution health risk assessment based on POI data. The method first acquires and preprocesses multi-source heterogeneous data to construct a POI spatial database; then, it uses Kriging spatial interpolation to generate a raster layer showing the spatial distribution of pollutant concentrations; next, it combines kernel density estimation and dynamic weight correction to construct a population activity intensity index surface; subsequently, it dynamically generates exposure assessment units to calculate the daily average exposure dose via multiple pathways for different age groups; finally, it quantifies the health risks via multiple pathways, generates a risk level distribution map, and identifies the dominant POI category in high-risk areas. This invention achieves deep fusion of multi-source data, accurately characterizes the correlation between population activity features and pollution exposure, improves the accuracy and adaptability of the assessment, and provides a scientific basis for pollution prevention and risk management.

Description

Technical Field

[0001] This invention relates to the field of regional soil pollution health risk assessment technology, specifically a method for regional soil pollution health risk assessment based on POI data. Background Technology

[0002] Current methods for assessing the health risks of regional soil pollution have many limitations. In terms of data utilization, traditional methods rely heavily on soil pollution sampling data and demographic data, resulting in relatively limited data sources. They fail to fully integrate various types of data that reflect the spatial characteristics of population activity, leading to an incomplete characterization of the relationship between population activity and pollution exposure.

[0003] In terms of the division of exposure assessment units, existing methods mostly use fixed grids or administrative regions as assessment units. This division method does not take into account the spatial differences in the intensity of population activity within the region and ignores the population agglomeration characteristics of different functional areas such as residential, educational, commercial and industrial areas. This results in a mismatch between the assessment units and the actual range of population activity, which in turn affects the accuracy of exposure dose calculation.

[0004] In characterizing population activity intensity, traditional methods often use static population density data, failing to incorporate the attraction effect of various locations on population activity and lacking a dynamic adjustment mechanism for the weight of population activity intensity. This makes it difficult to accurately reflect the spatiotemporal distribution characteristics of the population, resulting in discrepancies between health risk assessment results and actual conditions.

[0005] In addition, the results of existing assessment methods are presented in a relatively simple form, mainly in the form of data tables or simple charts, which makes it difficult to intuitively show the spatial distribution of high-risk areas and the causes of risks, causing inconvenience to relevant departments in formulating targeted pollution prevention and control measures and risk management strategies.

[0006] Therefore, there is an urgent need for a regional soil pollution health risk assessment method that can integrate multi-source data, accurately characterize population activity intensity, dynamically generate exposure assessment units, and achieve risk quantification and visualization, in order to solve the problems of insufficient accuracy, weak adaptability, and weak decision support of traditional assessment methods. Summary of the Invention

[0007] The purpose of this invention is to provide a method for assessing the health risks of regional soil pollution based on POI data, in order to solve the problems mentioned in the background art.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] A method for assessing the health risk of regional soil pollution based on POI data includes the following steps:

[0010] Step S1: Acquisition and preprocessing of multi-source heterogeneous data: Acquire soil pollution sampling point data, digital elevation model geospatial data, population census grid data, and Points of Interest (POI) data including categories such as residential, educational, medical, commercial, catering, park, and industrial facilities in the target area; classify, clean, geocode, and spatialize the POI data to establish a POI spatial database;

[0011] Step S2: Spatial distribution simulation of soil pollutants: Based on soil pollution sampling point data, a raster layer of the spatial distribution of specific pollutant concentrations within the target area is generated using the Kriging spatial interpolation method.

[0012] Step S3: Construction of dynamic weighted population activity intensity surface: Based on the POI spatial database, an initial activity intensity weight value is preset for each POI category; the kernel density estimation algorithm is used to generate density distribution raster layers for each type of POI; the initial weights of each type of POI are dynamically corrected according to the census grid data, and the corrected weights are weighted and superimposed with the corresponding density layers to generate a population activity intensity index surface that comprehensively reflects the spatiotemporal distribution characteristics of the population.

[0013] Step S4: Dynamic generation of exposure assessment units and calculation of pollution exposure dose: The population activity intensity index surface is divided into three levels of activity areas: high, medium and low, according to a preset threshold, and the spatial boundary of each level of activity area is used as a dynamic exposure assessment unit; for each exposure assessment unit, a concentration spatial distribution raster layer is superimposed, and the average pollutant concentration within the unit is extracted; combined with preset exposure parameters, the daily average pollutant exposure dose for different age groups within the unit is calculated for the three routes of oral ingestion, skin contact and inhalation.

[0014] Step S5: Multi-pathway health risk quantification and visualization: Substitute the pollutant exposure doses calculated in each exposure assessment unit and the corresponding pollutant toxicity parameters into the health risk assessment model to calculate the carcinogenic risk value and hazard quotient of each unit; map the calculated risk values ​​back to the corresponding dynamic exposure assessment unit spatial range to generate a regional soil pollution health risk level distribution map, and identify the dominant POI category information in the high-risk area.

[0015] As a preferred embodiment, step S1 specifically includes:

[0016] Soil pollution sampling point data is obtained through environmental monitoring networks, digital elevation model geospatial data and population census grid data are obtained through geographic information platforms, and raw point of interest (POI) data including categories such as residential, educational, medical, commercial, catering, park and industrial facilities are obtained through commercial map service interfaces or public datasets.

[0017] The original Points of Interest (POI) data is classified and cleaned. Based on the preset POI category system, the attribute information of each POI record is classified and filtered to remove duplicate, erroneous, and irrelevant POI records.

[0018] The cleaned POI records are geocoded and spatialized, converting the text address information of each POI record into precise geographic coordinates and unifying them to the same spatial reference coordinate system as the digital elevation model geospatial data and the census grid data.

[0019] Based on the spatialized POI records, geographic coordinates, and corresponding attribute category information, a structured POI spatial database is constructed. This database supports fast querying and statistical analysis based on spatial location and POI category, and provides an input data source for subsequent kernel density estimation algorithms.

[0020] As a preferred embodiment, step S2 specifically includes:

[0021] Based on soil pollution sampling point data, spatial structure analysis and outlier detection are performed to generate a standardized sampling point dataset that meets the requirements of spatial interpolation.

[0022] Based on a normalized sampling point dataset, a variogram model characterizing the spatial autocorrelation features of specific pollutant concentrations is constructed.

[0023] Based on the variogram model, the Kriging spatial interpolation algorithm is used to make the optimal unbiased estimate of the pollutant concentration at unknown spatial locations within the target area, and generate a spatial prediction point set of pollutant concentration.

[0024] The spatial prediction point set is rasterized to generate a preliminary raster layer of pollutant concentration spatial distribution with continuous concentration values ​​covering the entire target area.

[0025] Spatial reference and pixel attribute standardization processing are performed on the preliminary raster layer of pollutant concentration spatial distribution to generate a concentration spatial distribution raster layer that is spatially registered with the geospatial data of the digital elevation model and the subsequent surface spatial registration of the population activity intensity index.

[0026] As a preferred approach, based on a POI spatial database, an initial activity intensity weight value is preset for each POI category; a kernel density estimation algorithm is used to generate density distribution raster layers for each type of POI; specifically including:

[0027] Based on the POI spatial database, POI spatial points belonging to the categories of residential, educational, medical, commercial, catering, park and industrial facilities are extracted respectively;

[0028] For each POI category, an initial activity intensity weight value is assigned to characterize its relative importance based on its attraction and agglomeration effect on the daily activities of the population in the area.

[0029] For each category of POI spatial point set, a kernel density estimation algorithm is used independently for spatial smoothing to generate a density distribution raster layer that reflects the degree of spatial agglomeration of POI facilities of that type in the target area. The search radius of the kernel density estimation is set based on the average service range or spatial distribution characteristics of POI facilities of that type.

[0030] The generated density distribution raster layers of various POIs are normalized to ensure that their density values ​​are within a uniform range, thus forming standardized density distribution raster layers of various POIs.

[0031] As a preferred approach, the initial weights of various POIs are dynamically corrected based on census grid data. The corrected weights are then weighted and superimposed with the corresponding density layers to generate a population activity intensity index surface that comprehensively reflects the spatiotemporal distribution characteristics of the population. Specifically, this includes:

[0032] The standardized POI density distribution raster layers are spatially overlaid with census grid data to calculate the spatial correlation between the normalized density values ​​of various POIs in each grid cell and the number of permanent residents in the grid.

[0033] Based on the spatial correlation analysis results, a dynamic weight correction model is established. For POI categories that show a significant positive correlation with population size, their initial activity intensity weight values ​​are increased. For POI categories that show no significant correlation or a negative correlation, their initial weight values ​​are maintained or decreased, thus obtaining a set of dynamically corrected POI category activity intensity weight values.

[0034] The dynamically corrected POI category activity intensity weight values ​​are then multiplied by their corresponding standardized POI density distribution raster layers to generate weighted activity intensity raster layers for each category.

[0035] Spatially overlay and sum the weighted activity intensity raster layers for each category to generate a surface raster layer that comprehensively reflects the population activity intensity index caused by the combined effects of population distribution and multiple types of POI facilities.

[0036] As a preferred approach, the population activity intensity index surface is divided into high, medium, and low activity zones according to a preset threshold, and the spatial boundaries of each activity zone are used as dynamic exposure assessment units; specifically including:

[0037] Extract the cell values ​​of the surface raster layer of the population activity intensity index and calculate its statistical quantiles;

[0038] Based on statistical quantiles, two thresholds are set to divide the population activity areas into high, medium and low levels. The threshold for the high activity area is greater than or equal to the first high quantile, and the threshold for the medium activity area is between the first low quantile and the second low quantile.

[0039] Based on two thresholds, the surface raster layer of the population activity intensity index is reclassified to generate a population activity intensity partition raster layer with high, medium and low attributes.

[0040] The raster layer of population activity intensity is converted from raster to vector, and the continuous spatial boundaries of the high, medium and low activity zones are extracted to generate spatial vector boundary files as the basis for subsequent calculations.

[0041] The spatial vector boundary files are topologically checked and corrected to ensure that the boundaries of each level of activity area are non-overlapping and seamless. The corrected spatial vector boundaries of each level of activity area are defined as high, medium and low dynamic exposure assessment units, forming a dynamic exposure assessment unit vector layer.

[0042] As a preferred approach, for each exposure assessment unit, a concentration spatial distribution raster layer is overlaid to extract the average pollutant concentration within the unit. Combined with preset exposure parameters, the daily average pollutant exposure dose for different age groups within that unit is calculated via oral ingestion, skin contact, and inhalation. Specifically, this includes:

[0043] Spatial overlay analysis was performed between the dynamic exposure assessment unit vector layer and the concentration spatial distribution raster layer;

[0044] For each dynamic exposure assessment unit, the arithmetic mean of pollutant concentrations of all raster cells that fall completely within the boundary of the unit is calculated using a zonal statistical tool. This mean is defined as the representative average pollutant concentration of the dynamic exposure assessment unit.

[0045] Based on the activity zone level corresponding to the dynamic exposure assessment unit and the dominant POI category within it, and combined with the typical activity patterns of different age groups in POI locations of residential, educational, medical, commercial, catering, park and industrial facilities, a set of exposure parameters matching the population activity characteristics of the unit are determined and assigned. The exposure parameters include at least the daily stay time of different groups in each location, soil ingestion rate, skin exposure surface area and soil adhesion coefficient, and respiration rate.

[0046] The representative average pollutant concentration of the dynamic exposure assessment unit, the exposure parameters matched with the population activity characteristics of the unit, and the specific calculation formulas for different exposure routes were combined to independently calculate the average daily pollutant exposure dose for children and adults in the dynamic exposure assessment unit via oral ingestion, skin contact, and inhalation.

[0047] The daily average pollutant exposure doses calculated from three exposure pathways within each dynamic exposure assessment unit and for different population groups are aggregated to form a set of pollutant exposure dose results from multiple pathways and for multiple population groups, with the dynamic exposure assessment unit as the basic recording unit.

[0048] As a preferred approach, the pollutant exposure doses calculated within each exposure assessment unit, along with the corresponding pollutant toxicity parameters, are substituted into the health risk assessment model to calculate the carcinogenic risk value and hazard quotient for each unit; specifically including:

[0049] Based on the exposure dose results set of pollutants from multiple routes and multiple populations, the toxicity parameters of a specific pollutant are obtained. The toxicity parameters include at least the carcinogenicity slope factor of oral intake, the risk factor of inhalation unit, and the reference dose of each exposure route.

[0050] For each dynamic exposure assessment unit in the vector layer of the dynamic exposure assessment unit, the average daily oral exposure dose calculated within the unit for both children and adults is combined with the oral intake carcinogenic slope factor to calculate the carcinogenic risk value of the oral intake route for that unit.

[0051] For each dynamic exposure assessment unit in the vector layer of the dynamic exposure assessment unit, the average daily exposure dose of respiratory inhalation calculated within the unit for both children and adults is combined with the risk factor of the respiratory inhalation unit to calculate the carcinogenic risk value of the respiratory inhalation route for that unit.

[0052] For each dynamic exposure assessment unit in the vector layer of the dynamic exposure assessment unit, the average daily exposure doses calculated within the unit for both children and adults for oral ingestion, skin contact, and inhalation are combined with the reference doses for the corresponding exposure routes to calculate the hazard quotient for each route.

[0053] For each dynamic exposure assessment unit, the carcinogenic risk values ​​of oral ingestion and inhalation are summarized to obtain the comprehensive carcinogenic risk value of the unit; the hazard quotients of oral ingestion, skin contact and inhalation are summarized to obtain the comprehensive hazard quotient index of the unit.

[0054] Based on the comprehensive carcinogenic risk value and comprehensive hazard quotient of all dynamic exposure assessment units, a set of regional soil pollution health risk assessment results is constructed.

[0055] As a preferred approach, the calculated risk values ​​are mapped back to the corresponding spatial range of dynamic exposure assessment units to generate a regional soil pollution health risk level distribution map, and the dominant POI category information within high-risk areas is associated for identification; specifically including:

[0056] Based on the regional soil pollution health risk assessment results set, and according to the preset risk level classification thresholds of comprehensive carcinogenic risk value and comprehensive hazard quotient index, the risk level of each dynamic exposure assessment unit is assessed.

[0057] The assessed risk level attributes are associated with the corresponding spatial elements in the dynamic exposure assessment unit vector layer to generate a dynamic exposure assessment unit risk zoning vector layer with risk level attributes.

[0058] Symbolic rendering is performed on the vector map layer of risk zoning of dynamic exposure assessment units, and different colors and legends are assigned according to the risk level to generate a regional soil pollution health risk level distribution map.

[0059] For dynamic exposure assessment units that are rated as high-risk in the risk zoning vector layer of dynamic exposure assessment units, their spatial range is overlaid with the POI spatial database to identify and extract the dominant POI category within the spatial range of the unit.

[0060] In the regional soil pollution health risk level distribution map, high-risk dynamic exposure assessment units are marked, and their dominant POI category information is used as auxiliary information for risk causes for association identification.

[0061] As can be seen from the technical solution provided by the present invention above, the regional soil pollution health risk assessment method based on POI data provided by the present invention has the following beneficial effects:

[0062] Significantly improved assessment accuracy: This invention integrates soil pollution sampling point data, digital elevation model geospatial data, population census grid data, and multi-category POI data to achieve deep fusion of multi-source heterogeneous data, breaking the limitations of traditional single assessment data; through POI data classification, cleaning, geocoding, and spatialization processing, combined with kernel density estimation algorithm and dynamic weight correction model, it accurately depicts the spatiotemporal distribution characteristics of population activity intensity, making pollution exposure assessment highly consistent with actual population activity patterns, and greatly improving the accuracy and reliability of health risk assessment results;

[0063] The dynamic assessment mechanism is highly adaptable: It innovatively adopts a dynamic exposure assessment unit generation method, dividing high, medium and low activity zones based on the statistical quantiles of the population activity intensity index surface. The spatial boundary of the activity zone is used as the assessment unit, replacing the traditional fixed grid unit, which better reflects the spatial differences in regional population activities. At the same time, it matches differentiated exposure parameters to the typical activity patterns of different age groups in various POI locations, and calculates the daily average exposure dose through multiple pathways. This ensures that the assessment process fully considers population heterogeneity, adapts to the assessment needs of different regions and scenarios, and has a wider range of applications.

[0064] The system demonstrates outstanding risk management support capabilities: through multi-channel health risk quantification calculations, it generates comprehensive carcinogenic risk values ​​and comprehensive hazard quotient indices, and combines them with symbolic rendering technology to form an intuitive regional soil pollution health risk level distribution map. At the same time, it identifies the dominant POI category information in high-risk areas, clearly revealing the spatial distribution and causal relationships of high-risk areas. This provides relevant departments with clear data support and decision-making basis for accurately locating key pollution prevention and control areas, formulating targeted control measures, and optimizing regional land use planning, significantly improving the scientific nature and efficiency of risk management. Attached Figure Description

[0065] Figure 1 This is a schematic diagram of the steps in a regional soil pollution health risk assessment method based on POI data according to the present invention. Detailed Implementation

[0066] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0067] To better understand the above technical solutions, the following will provide a detailed description of the technical solutions in conjunction with the accompanying drawings and specific embodiments.

[0068] like Figure 1 As shown, this embodiment of the invention provides a method for assessing regional soil pollution health risks based on POI data, including the following steps:

[0069] Step S1: Acquisition and preprocessing of multi-source heterogeneous data: Acquire soil pollution sampling point data, digital elevation model geospatial data, population census grid data, and Points of Interest (POI) data including categories such as residential, educational, medical, commercial, catering, park, and industrial facilities in the target area; classify, clean, geocode, and spatialize the POI data to establish a POI spatial database;

[0070] Step S2: Spatial distribution simulation of soil pollutants: Based on soil pollution sampling point data, a raster layer of the spatial distribution of specific pollutant concentrations within the target area is generated using the Kriging spatial interpolation method.

[0071] Step S3: Construction of dynamic weighted population activity intensity surface: Based on the POI spatial database, an initial activity intensity weight value is preset for each POI category; the kernel density estimation algorithm is used to generate density distribution raster layers for each type of POI; the initial weights of each type of POI are dynamically corrected according to the census grid data, and the corrected weights are weighted and superimposed with the corresponding density layers to generate a population activity intensity index surface that comprehensively reflects the spatiotemporal distribution characteristics of the population.

[0072] Step S4: Dynamic generation of exposure assessment units and calculation of pollution exposure dose: The population activity intensity index surface is divided into three levels of activity areas: high, medium and low, according to a preset threshold, and the spatial boundary of each level of activity area is used as a dynamic exposure assessment unit; for each exposure assessment unit, a concentration spatial distribution raster layer is superimposed, and the average pollutant concentration within the unit is extracted; combined with preset exposure parameters, the daily average pollutant exposure dose for different age groups within the unit is calculated for the three routes of oral ingestion, skin contact and inhalation.

[0073] Step S5: Multi-pathway health risk quantification and visualization: Substitute the pollutant exposure doses calculated in each exposure assessment unit and the corresponding pollutant toxicity parameters into the health risk assessment model to calculate the carcinogenic risk value and hazard quotient of each unit; map the calculated risk values ​​back to the corresponding dynamic exposure assessment unit spatial range to generate a regional soil pollution health risk level distribution map, and identify the dominant POI category information in the high-risk area.

[0074] In this embodiment, step S1 aims to comprehensively collect various types of soil pollution-related data from the target area. Through classification, cleaning, geocoding, and spatialization processing, a standardized and structured POI spatial database is constructed, providing complete and accurate basic data support for subsequent spatial distribution simulation of soil pollutants, population activity intensity analysis, and risk assessment. The detailed steps are as follows:

[0075] Step S1-1: Comprehensive Acquisition of Multi-Source Heterogeneous Data

[0076] Based on the core needs of regional soil pollution health risk assessment, four types of core data are acquired through multiple channels: The first type is soil pollution sampling point data, collected through the regional environmental monitoring network, including key information such as the geographic coordinates of the sampling points, sampling depth, types of monitored pollutants, and corresponding concentration values. The second type is digital elevation model geospatial data, acquired through a professional geographic information platform, covering topographic and geomorphic spatial information such as elevation, slope, and aspect of the target area. The third type is population census grid data, also sourced from the geographic information platform, using standardized grids as units, recording population distribution data such as the number of permanent residents and population density within each grid. The fourth type is multi-category Point of Interest (POI) data, collected through commercial map service interfaces or public datasets, covering seven core categories: residential, educational, medical, commercial, catering, park, and industrial facilities. Each data entry includes attribute information such as POI name, address, and category. During data acquisition, the system automatically records metadata information such as collection time and data accuracy of each data source to ensure data traceability.

[0077] Step S1-2: POI data classification and cleaning process:

[0078] Based on a pre-defined POI category system, the collected raw POI data is classified and cleaned. First, according to the categorization standards for residential, educational, medical, commercial, catering, park, and industrial facilities, the attribute information of each raw POI record is compared and categorized one by one to clarify the target category of each record. Then, the data cleaning process is initiated, using a duplicate record identification algorithm to remove POI records with completely identical content or duplicate core attributes. Error records are filtered based on address format validity verification rules, including records with missing or logically contradictory address information. POI records unrelated to the target area are filtered according to the assessment scope boundaries, retaining only valid records located within the assessment area or closely related to the area's population activities. After cleaning, a clearly categorized, redundant, and error-free intermediate dataset of POIs is generated, and a data cleaning report is output, specifying the total amount of raw data, the number of records removed, and the number of valid records.

[0079] Steps S1-3: Geocoding and Spatialization of POI Data:

[0080] Geocoding and spatialization transformation are performed on the cleaned intermediate POI dataset. First, a geocoding tool is used to convert the text address information of each POI record into precise latitude and longitude geographic coordinates, ensuring that the coordinate accuracy meets the requirements of spatial analysis. Then, spatial reference coordinate system one processing is performed to uniformly project the transformed POI geographic coordinates, digital elevation model geospatial data, and census grid data onto the same preset spatial reference coordinate system, eliminating spatial benchmark differences between different data sources. During the coordinate transformation and coordinate system one process, the system automatically verifies the transformation accuracy. If the coordinate transformation error of a single POI exceeds a preset threshold, it is marked as an abnormal record and recoded. If the overall data coordinate system transformation consistency does not meet the standard, the transformation process is re-executed to ensure that the spatial benchmark of all spatial data is unified and accurately matched.

[0081] Steps S1-4: Structured Construction of the POI Spatial Database

[0082] Based on the core information from spatialized POI data, digital elevation model geospatial data, and census grid data, a structured POI spatial database is constructed. The database comprises a spatial data layer and an attribute data layer. The spatial data layer stores the coordinate information and spatial topological relationships of POIs and other geographic data. The attribute data layer stores the category, name, address, and other attribute information of POIs, as well as the terrain attributes of the digital elevation model and the population attribute data of the census grid. This database supports two core query and statistical functions: first, location-based query analysis, which can quickly locate all POIs and their corresponding geographic and population data within a specific spatial range; second, category-based filtering and statistics, which can batch extract the spatial distribution and attribute information of POIs of a specific category. After the database is constructed, automatic data integrity and query response speed tests are performed to ensure no data loss and efficient query and statistical analysis, providing a stable input data source for subsequent kernel density estimation algorithms and other spatial analysis processes.

[0083] In this embodiment, step S2 aims to generate a spatially continuous raster layer of pollutant concentration distribution based on soil pollution sampling point data in the target area. This layer is achieved through data normalization, variogram modeling, kriging spatial interpolation calculation, and layer standardization optimization. The layer is fully covered, spatially continuous, and registered with other geographic data, providing accurate spatial data support for pollutant concentration calculation in subsequent exposure assessment units. The detailed steps are as follows:

[0084] Step S2-1: Normalization of sampling point data:

[0085] Based on the collected soil pollution sampling data, spatial structure analysis and outlier detection were conducted to generate a standardized sampling point dataset that meets the requirements of spatial interpolation. First, spatial autocorrelation analysis was used to explore the spatial distribution pattern of the sampling point data, determine the clustering, dispersion, or random distribution characteristics of pollutant concentrations in space, and clarify the rationality of the data's spatial structure. Then, an outlier detection process was initiated, using the Grubbs test to verify the pollutant concentration values ​​at each sampling point, identifying and eliminating extreme outliers caused by sampling errors, instrument malfunctions, or abnormal pollution events. After outlier detection, the remaining valid sampling point data were integrated, and key attribute information such as geographic coordinates, pollutant type, and concentration values ​​for each sampling point were supplemented and improved to form a standardized sampling point dataset with a complete structure and uniform format, ensuring that the data meets the basic requirements for subsequent spatial interpolation modeling.

[0086] Step S2-2: Construction of the variogram model:

[0087] Based on a normalized sampling point dataset, a variogram model characterizing the spatial autocorrelation features of specific pollutant concentrations is constructed. The variogram is a core tool for describing the spatial variation patterns of pollutant concentrations, and its calculation formula is as follows:

[0088] ,in, The spacing is The value of the variogram. Spacing equals Number of sampling points For the first The spatial location of each sampling point for Pollutant concentration at the location To and Position spacing is The pollutant concentration values ​​at the sampling points were calculated. During the calculation process, firstly, reasonable distance grouping intervals were determined based on the spatial coordinates of the standardized sampling point dataset. Then, the number of sampling point pairs and the corresponding sum of squared concentration differences within each distance interval were counted and substituted into the formula to calculate the pollutant concentration values ​​at different intervals. The corresponding variogram values ​​were then used to construct a theoretical variogram model that conforms to the spatial variation characteristics of pollutant concentrations, based on these discrete variogram values. Key parameters such as the nugget sill value and range of the model were then determined, providing model support for subsequent Kriging space interpolation.

[0089] Step S2-3: Kriging space interpolation calculation:

[0090] Based on the constructed variogram model, the Kriging spatial interpolation algorithm is used to perform optimal unbiased estimation of pollutant concentrations at unknown spatial locations within the target area, generating a spatial prediction point set for pollutant concentrations. First, the target area is divided into spatial prediction grids, with the grid size reasonably set according to the evaluation accuracy requirements and the area's scope, ensuring that the prediction results meet accuracy needs without excessively increasing computational load. For the center point of each prediction grid, the Kriging interpolation algorithm is used to calculate the optimal unbiased estimate of the pollutant concentration at that center point based on its spatial positional relationship with all surrounding sampling points, combined with the spatial variability information provided by the variogram model. During the calculation, the algorithm automatically assigns different weights to sampling points at different distances; closer sampling points have higher weights, and farther sampling points have lower weights, ensuring that the estimation results fully reflect the spatial correlation of pollutant concentrations. After estimating the concentration at the center points of all prediction grids one by one, all prediction results are integrated to form a spatial prediction point set for pollutant concentrations covering all unknown spatial locations within the target area.

[0091] Step S2-4: Initial raster layer generation:

[0092] The obtained spatial prediction point set of pollutant concentration is rasterized to generate a preliminary raster layer of pollutant concentration spatial distribution with continuous concentration values ​​covering the entire target area. During the rasterization process, based on the prediction grid set in step S2-3, the estimated pollutant concentration value of the center point of each prediction grid is assigned to the corresponding raster cell, so that the discrete prediction point set is transformed into a continuous raster data structure. At the same time, the spatial range of the raster layer is set to be completely consistent with the target area to ensure that the layer can completely cover the evaluation area without any missing data or exceeding the range. In the generated preliminary raster layer, the value of each cell represents the estimated pollutant concentration value at that spatial location. Through the continuous distribution of cell values, the spatial distribution trend of pollutants in the target area is intuitively presented.

[0093] Step S2-5: Layer Normalization and Spatial Registration:

[0094] Spatial reference and pixel attribute standardization processes were performed on the initial raster layer of pollutant concentration spatial distribution to generate a concentration spatial distribution raster layer spatially registered with the digital elevation model geospatial data and the subsequent population activity intensity index surface. For spatial reference standardization, the spatial reference coordinate system of the initial raster layer was uniformly adjusted to the same coordinate system as the population census grid data in the digital elevation model geospatial data, ensuring that all geographic data are spatially consistent and avoiding deviations in subsequent overlay analysis due to coordinate system differences. For pixel attribute standardization, the pixel size data format and concentration value unit of the raster layer were unified, and the rationality of the concentration values ​​was verified, eliminating possible outliers to ensure that the pixel attributes meet the requirements of subsequent data processing and analysis. After standardization, the final concentration spatial distribution raster layer was output. This layer not only has continuous and accurate pollutant concentration information but also achieves precise spatial registration with other core geographic data, laying the foundation for concentration extraction and dose calculation in subsequent exposure assessment units.

[0095] In this embodiment, step S3 is to construct a population activity intensity index surface that comprehensively reflects the spatiotemporal distribution characteristics of the population based on the POI spatial database and census grid data, through POI category extraction, initial weight preset, kernel density estimation, dynamic weight correction, and weighted overlay, providing accurate spatial data support for the subsequent division of dynamic exposure assessment units; the detailed steps are as follows:

[0096] Step S3-1: POI category extraction and initial activity intensity weight preset:

[0097] Based on the POI spatial database, POI locations belonging to seven categories—residential, educational, medical, commercial, catering, park, and industrial facilities—are extracted through category filtering to ensure that the attribute category of each POI location fully matches the preset category system. Then, based on the attraction and agglomeration effect of each type of POI facility on the daily activities of the regional population, initial activity intensity weights representing their relative importance are assigned. For example, residential POIs are given higher initial weights due to their long dwell time and strong agglomeration effect, industrial facility POIs are given moderate initial weights due to their relatively low population activity frequency, and park POIs are given corresponding initial weights based on their opening hours and the size of their user base. After the weight assignment is completed, a correlation table between POI categories and initial weights is established to ensure that the initial weights of each POI category are traceable and adjustable, providing basic data for subsequent dynamic correction.

[0098] Step S3-2: POI kernel density estimation and density distribution raster layer generation:

[0099] For each category of POI spatial location set, a kernel density estimation algorithm is independently initiated for spatial smoothing. First, based on the average service range or spatial distribution characteristics of each type of POI facility, a search radius for kernel density estimation is set. For example, the search radius for residential POIs is set based on the average service radius of urban residential areas, while the search radius for commercial POIs is determined by referring to the radiation range of commercial complexes. Then, the kernel density estimation algorithm calculates the POI clustering degree at each spatial location, using the following formula:

[0100] ,in, For spatial location The kernel density value at that location, This represents the total number of spatial points in this category of POIs. The set search radius, For the first The coordinates of each POI spatial point are obtained; the spatial location within the target area is calculated one by one using this formula to generate a density distribution raster layer that reflects the spatial agglomeration degree of this type of POI facility; finally, the generated POI density distribution raster layers are normalized by using a linear normalization method to convert all density values ​​to a unified numerical range of 0 to 1, forming standardized POI density distribution raster layers for various types of POIs and eliminating the dimensional differences in density values ​​of different types of POIs.

[0101] Step S3-3: Dynamic weight correction and generation of weighted activity intensity raster layer:

[0102] This paper describes a spatial overlay analysis of standardized POI density distribution raster layers and census grid data. Each grid cell is associated with both the normalized density value of various POIs and the number of permanent residents within the grid. The spatial correlation coefficient between the normalized density value of various POIs and the number of permanent residents within each grid cell is then calculated to quantify the degree of association. Based on the spatial correlation analysis results, a dynamic weight correction model is established. For POI categories with a correlation coefficient absolute value greater than a preset threshold and exhibiting a significant positive correlation, their initial activity intensity weight value is increased by a preset ratio. For POI categories with a correlation coefficient absolute value less than the preset threshold and exhibiting no significant correlation or a negative correlation, their initial weight value is maintained or decreased by a preset ratio. After completing the weight correction for all POI categories, a set of dynamically corrected POI category activity intensity weight values ​​is obtained. These dynamically corrected weight values ​​are then multiplied by their corresponding standardized POI density distribution raster layers, i.e., the density value of each pixel is multiplied by the corrected weight of the corresponding category, generating weighted activity intensity raster layers for each category. This allows the layers to simultaneously reflect the clustering degree and weight priority of POIs.

[0103] Step S3-4: Surface integration generation of population activity intensity index:

[0104] Spatial overlay and summation operations are performed on the weighted activity intensity raster layers of each category. The weighted activity intensity pixel values ​​of each category corresponding to each spatial location within the target area are accumulated to generate a comprehensive raster layer. Each pixel value in this layer is the population activity intensity index of the corresponding spatial location. Its value comprehensively reflects the population distribution and the density of population activity under the combined effect of multiple types of POI facilities. After generation, the raster layer is spatially referenced to ensure that it is spatially registered with the digital elevation model geospatial data population census grid data and the concentration spatial distribution raster layer to be generated subsequently. The final population activity intensity index surface raster layer provides core population activity spatial distribution data support for the division of dynamic exposure assessment units in step S4.

[0105] In this embodiment, step S4 is based on the population activity intensity index surface generated in step S3 and the concentration spatial distribution raster layer generated in step S2. Through activity area classification, dynamic exposure assessment unit construction, extraction of average pollutant concentration within the unit, and multi-pathway, multi-population exposure dose calculation, a standardized exposure dose result set is formed, providing accurate exposure data support for the health risk quantification in step S5. The detailed steps are as follows:

[0106] Step S4-1: Population activity intensity pixel value statistics and threshold setting:

[0107] All pixel values ​​in the surface raster layer of the population activity intensity index are extracted to construct a pixel value dataset. Statistical analysis methods are used to calculate the statistical quantiles of this dataset, including the first upper quantile, the second lower quantile, and the first lower quantile. The first upper quantile is typically taken as the 75th quantile, the second lower quantile as the 50th quantile, and the first lower quantile as the 25th quantile. Based on these statistical quantiles, two thresholds are set to divide the population activity areas into high, medium, and low levels. The first threshold is the high activity area threshold, which is greater than or equal to the first upper quantile. The second threshold is the medium activity area threshold, which is between the first and second lower quantiles. Pixel values ​​below the first lower quantile are designated as low activity areas. After setting the thresholds, a threshold division standard table is generated to clarify the pixel value range corresponding to each level of activity area, providing a basis for subsequent pixel reclassification.

[0108] Step S4-2: Generation of raster layer for population activity intensity zoning:

[0109] Based on two set thresholds, a cell reclassification operation is performed on the surface raster layer of the population activity intensity index. Cells with values ​​greater than or equal to the high activity zone threshold are assigned the high activity zone attribute, cells with values ​​between the medium and high activity zone thresholds are assigned the medium activity zone attribute, and cells with values ​​lower than the medium activity zone threshold are assigned the low activity zone attribute. During the reclassification process, the system automatically verifies the accuracy of the attribute assignment for each cell to ensure that no cells are missed or misclassified. After the reclassification is completed, a population activity intensity zoning raster layer with high, medium, and low attribute labels is generated. This layer visually presents the spatial hierarchical characteristics of population activity intensity within the target area.

[0110] Step S4-3: Construction of the vector layer for dynamic exposure assessment unit:

[0111] The raster-to-vector conversion tool was used to convert the raster layer representing population activity intensity zones. During the conversion, based on the attributes of each activity zone, continuous spatial boundaries of high, medium, and low activity zones were extracted to generate corresponding spatial vector boundary files. Each file contained spatial information such as boundary coordinates and topological relationships for the corresponding activity zone level. The generated spatial vector boundary files underwent a topology check, focusing on verifying whether there were any topological errors such as overlapping, gaps, or hanging nodes. For any errors found, a topology correction tool was used to repair them, ensuring that the boundaries of each activity zone were free of overlap and gaps and that the topological relationships were correct. The corrected spatial vector boundaries of the high, medium, and low activity zones were defined as high, medium, and low dynamic exposure assessment units, respectively, and integrated to form a dynamic exposure assessment unit vector layer. This layer provides the spatial unit basis for subsequent concentration extraction and dose calculation.

[0112] Step S4-4: Extraction of average pollutant concentration from the exposure assessment unit:

[0113] Spatial overlay analysis was performed on the vector layer of the dynamic exposure assessment unit and the raster layer of concentration spatial distribution to establish their spatial correlation. For each dynamic exposure assessment unit, a zoning statistical tool was used to select all raster pixels that fall completely within the unit's boundary range, and the pollutant concentration values ​​corresponding to these pixels were extracted. The average value of these concentration values ​​was calculated using the arithmetic mean method, with the following formula:

[0114] ,in, The representative average pollutant concentration for the dynamic exposure assessment unit. The total number of raster cells that completely fall within this cell. For the first The pollutant concentration value of each raster cell; the calculated As the core concentration parameter of this dynamic exposure assessment unit, it is associated and stored in the attribute table of the dynamic exposure assessment unit vector layer to form a comprehensive dataset containing spatial and concentration information.

[0115] Step S4-5: Exposure parameter matching and calculation of daily average exposure dose from multiple routes and populations:

[0116] Based on the activity zone level corresponding to the dynamic exposure assessment unit and the dominant POI category within it, and combined with the typical activity patterns of different age groups in various POI locations, matching exposure parameters are determined and assigned. The exposure parameters include the daily stay time, soil ingestion rate, skin exposed surface area, soil adhesion coefficient, and respiration rate of different groups in each location.

[0117] Authoritative sources for exposure parameters:

[0118] All exposure parameters should prioritize the nationally recommended values ​​in the "Handbook of Exposure Parameters for Chinese Populations" (Children's Volume and Adult Volume) issued by the Ministry of Ecology and Environment, and can be finely adjusted locally based on the population activity characteristics of the target area; for parameters without domestic recommended values, the recommended values ​​in the "Technical Guidelines for Risk Assessment of Soil Pollution in Construction Land" (HJ25.3-2019) should be adopted to ensure the compliance, authority and reproducibility of the parameters;

[0119] Definition of core parameters, basis for assigning differentiated values ​​to different population groups, and recommended value ranges:

[0120] Daily stay time This refers to the daily duration of a person's stay within the corresponding PO type assessment unit, measured in hours (h / d). Differential assignments for children and adults are based on the daily activity patterns of different groups; for example, in residential POI units, adults... The recommended value is 16 hours / day for children. Recommended dose: 20 hours / day; Commercial PO1 unit, adult. The recommended value is 2 hours / day for children. The recommended value is 1h / d, and the specific value should be determined based on the dominant POI category within the cell.

[0121] Soil intake rate This refers to the amount of soil ingested orally by a population each day, measured in units of... According to the "Chinese Population Exposure Parameter Manual", the recommended value for children is... Recommended value for adults ;

[0122] Exposed surface area of ​​skin : refers to the skin surface area in contact with soil, measured in units of According to the "Chinese Population Exposure Parameter Manual", the recommended value for children is... Recommended value for adults ;

[0123] Soil adhesion coefficient This refers to the amount of soil adhering to a unit area of ​​skin surface, measured in units of... According to HJ25.3-2019, the recommended value for children is... Recommended value for adults ;

[0124] Skin absorption coefficient This refers to the proportion of pollutants absorbed through the skin, and is dimensionless. Based on HJ25.3-2019, values ​​are assigned for different types of pollution; the recommended value for heavy metal pollutants is... The recommended value for volatile organic compounds is determined by combining the octanol-water partition coefficient of the pollutant;

[0125] respiratory rate This refers to the volume of air a person breathes daily, measured in units of... According to the "Chinese Population Exposure Parameter Manual", the recommended value for children is... Recommended value for adults ;

[0126] Exposure frequency This refers to the number of days per year that a person is exposed to soil pollutants, expressed in units of... The general recommended value for regional-scale assessment is ;

[0127] Exposure duration This refers to the number of years a person is continuously exposed to soil pollutants, measured in units of... In the carcinogenic risk assessment, adults Recommended value ,child Recommended value In the non-carcinogenic risk assessment, The value is equal to the exposure duration for the corresponding population.

[0128] weight This refers to the average weight of the corresponding population, in units of... According to the "Chinese Population Exposure Parameter Manual", the recommended value for children is... Recommended value for adults ;

[0129] Mean exposure time : refers to the average exposure time to pollutants, in units of In the carcinogenic risk assessment, Values (Average life expectancy of the corresponding population); In the non-carcinogenic risk assessment, Values ;

[0130] Parameter matching rules:

[0131] For dynamic exposure assessment units with high, medium, and low activity intensities, the daily stay time is assessed in conjunction with the dominant POI category within the unit. Differentiated values ​​were assigned to the parameters, while the other parameters adopted the authoritative recommended values ​​mentioned above, to ensure that the assessment results closely matched the actual exposure characteristics of the population in the region.

[0132] The representative average pollutant concentration of the dynamic exposure assessment unit, the matched exposure parameters, and the specific calculation formulas for different exposure routes are combined to calculate the daily average exposure dose for oral ingestion, skin contact, and inhalation.

[0133] The formula for calculating the average daily exposure dose via oral intake is:

[0134] ,in, This represents the average daily exposure dose via the oral route. For soil uptake rate, For daily stay time, For exposure frequency, For the duration of exposure, For weight, The average exposure time is used; the formula and parameter definitions comply with the requirements of the "Technical Guidelines for Soil Pollution Risk Assessment of Construction Land" (HJ25.3-2019).

[0135] The formula for calculating the average daily exposure dose via skin contact is:

[0136] ,in, This refers to the average daily exposure dose via skin contact. The exposed surface area of ​​the skin. Soil adhesion coefficient, This is the skin absorption coefficient; the meanings of the other symbols are consistent with those in the formula for oral ingestion. This formula and parameter definitions comply with the requirements of the "Technical Guidelines for Risk Assessment of Soil Pollution in Construction Land" (HJ25.3-2019).

[0137] The formula for calculating the average daily exposure dose via the inhalation route is:

[0138] ,in, This refers to the average daily exposure dose via the inhalation route. The respiratory rate is given by the formula; the meanings of the other symbols are consistent with those for the oral intake route. This formula and parameter definitions comply with the requirements of the "Technical Guidelines for Risk Assessment of Soil Pollution in Construction Land" (HJ25.3-2019).

[0139] After calculating the average daily exposure dose for each of the three pathways separately for children and adults, the exposure dose data of all population groups within each dynamic exposure assessment unit are summarized to form a multi-pathway and multi-population exposure dose result set of pollutants with the dynamic exposure assessment unit as the basic recording unit. This result set includes key information such as unit identifier, population type, exposure dose of each pathway, and total exposure dose.

[0140] In this embodiment, step S5 is based on the multi-pathway and multi-population exposure dose result set of pollutants generated in step S4. Combined with the toxicity parameters of specific pollutants, it quantifies and intuitively displays the health risks of regional soil pollution through carcinogenic risk value and hazard quotient calculation, risk level assessment, visualization, and identification of the causes of high-risk areas. This provides a precise decision-making basis for pollution prevention and risk management. The detailed steps are as follows:

[0141] Step S5-1: Acquisition and correlation of toxicity parameters:

[0142] Based on the multi-pathway and multi-population exposure dose result set of pollutants, the specific types of pollutants involved in the assessment are clearly identified, and the toxicity parameters of the corresponding pollutants are obtained from the pre-set toxicity parameter database. The toxicity parameters include at least the carcinogenicity slope factor of oral intake, the risk factor of inhalation unit, and the reference dose of oral intake, skin contact, and inhalation.

[0143] Authoritative sources of core toxicity parameters:

[0144] Oral intake of carcinogenic slope factors The recommended pollutant values ​​published in the appendices of the "Soil Environmental Quality Standard for Construction Land Soil Pollution Risk Control" (GB 15618-2018) and the "Technical Guidelines for Soil Pollution Risk Assessment of Construction Land" (HJ 25.3-2019) shall be given priority. If there are no recommended values ​​in domestic standards, the authoritative toxicity parameter values ​​published by the US EPA Integrated Risk Information System (IRIS) shall be adopted.

[0145] Inhalation risk factors The recommended pollutant values ​​published in the appendix of the "Technical Guidelines for Soil Pollution Risk Assessment of Construction Land" (HJ 25.3-2019) shall be given priority; if no recommended values ​​are available in domestic standards, the recommended unit risk factor values ​​published by the Integrated Risk Information System (IRIS) of the U.S. EPA shall be adopted.

[0146] Reference values ​​for each exposure route , , Priority should be given to the domestic recommended values ​​published in the "Technical Guidelines for Risk Assessment of Soil Pollution in Construction Land" (HJ 25.3-2019). If no domestic recommended values ​​are available, authoritative values ​​from the US EPA IRIS database should be used.

[0147] Explanation of differences among population groups:

[0148] The above-mentioned toxicity parameters are inherent properties of pollutants, determined by the physicochemical and toxicological characteristics of the pollutants themselves, and do not change with the age of the population; the difference in risk values ​​between children and adults is achieved through the calculation of differentiated exposure doses for the two groups, and is unrelated to the toxicity parameters themselves, and there is no population-differentiated toxicity parameter setting.

[0149] During the acquisition process, the system automatically verifies the applicability and timeliness of toxicity parameters to ensure that the parameters are fully matched with the types of pollutants and comply with the current industry standards or specifications for environmental health risk assessment. The acquired toxicity parameters are associated and stored with the results of exposure doses of pollutants through multiple pathways and among multiple populations. A comprehensive dataset is established with dynamic exposure assessment units as the core index, which associates pollutant types, exposure doses and corresponding toxicity parameters, providing complete parameter support for subsequent risk calculations.

[0150] Step S5-2: Calculation of carcinogenic risk value via multiple pathways:

[0151] For each dynamic exposure assessment unit in the vector layer of the dynamic exposure assessment unit, the carcinogenic risk values ​​for oral ingestion, skin contact and inhalation are calculated based on the associated comprehensive dataset.

[0152] The formula for calculating the carcinogenic risk value via oral intake is:

[0153] ,in, This represents the carcinogenic risk value via oral intake. This represents the average daily exposure dose via oral ingestion. For oral intake of carcinogenic slope factors;

[0154] The formula for calculating the carcinogenic risk value via skin contact is:

[0155] ,in, This represents the carcinogenic risk value via skin contact. This refers to the average daily exposure dose through skin contact. A carcinogenic slope factor that can be detected through skin contact;

[0156] The formula for calculating the carcinogenic risk value via inhalation is as follows:

[0157] ,in, This represents the carcinogenic risk value via inhalation. This refers to the average daily exposure dose via inhalation. Risk factors for inhaled respiratory units;

[0158] After independently calculating the carcinogenic risk values ​​for the two pathways for both children and adults, the carcinogenic risk values ​​for both pathways for each dynamic exposure assessment unit are summarized to obtain the comprehensive carcinogenic risk value for that unit. The formula is as follows:

[0159] ,in, The comprehensive carcinogenic risk value for the dynamic exposure assessment unit. The risk value for carcinogenicity in children via oral intake. This represents the carcinogenic risk value for adults via oral intake. For children's skin contact carcinogenic risk values, This represents the carcinogenic risk value for adults via skin contact. For children, the risk value of carcinogenicity via inhalation is... For adults, the carcinogenic risk value is the value for carcinogenicity via inhalation.

[0160] Step S5-3: Calculation of Multi-Pathway Hazard Quotient and Summary of Composite Index:

[0161] For each dynamic exposure assessment unit in the vector layer of the dynamic exposure assessment unit, the hazard quotients for oral ingestion, skin contact and inhalation are calculated based on the associated comprehensive dataset.

[0162] The formula for calculating the hazard quotient via oral intake is:

[0163] ,in, For the harm caused by oral ingestion, This represents the average daily exposure dose via oral ingestion. This is a reference dose for oral intake;

[0164] The formula for calculating the hazard quotient via skin contact is:

[0165] ,in, For products that pose a risk through skin contact, This refers to the average daily exposure dose through skin contact. For skin contact reference dose;

[0166] The formula for calculating the hazard quotient via inhalation is:

[0167] ,in, For hazards posed by inhalation, This refers to the average daily exposure dose via inhalation. This is the reference dose for inhalation.

[0168] After calculating the hazard quotients for the three pathways separately for children and adults, the hazard quotients for the three pathways for each dynamic exposure assessment unit are summarized for both groups to obtain the comprehensive hazard quotient index for that unit. The formula is as follows:

[0169] ,in, The comprehensive hazard quotient index for dynamic exposure assessment units. The risk of children ingesting these substances orally. For adults, the risk factors for harm via oral ingestion are as follows: For children's skin contact hazards, For adults, the risk of skin contact hazard is... For children, the risk of respiratory inhalation hazards Hazards associated with the adult respiratory inhalation route;

[0170] Step S5-4: Construction of the regional soil pollution health risk assessment result set:

[0171] Based on the comprehensive carcinogenic risk value and comprehensive hazard quotient of all dynamic exposure assessment units, the system integrates the unit identifier, spatial coordinates, population exposure dose, risk values ​​of each pathway, and comprehensive risk indicators of each unit to construct a regional soil pollution health risk assessment result set. This result set is stored in a structured data format and supports multi-dimensional querying and statistical analysis by dynamic exposure assessment unit, risk level, POI category, etc. After construction, the system automatically performs data integrity verification to ensure that the risk calculation results of each dynamic exposure assessment unit are complete and without anomalies, providing a reliable data foundation for subsequent risk level assessment and visualization.

[0172] Step S5-5: Risk Level Assessment and Risk Zoning Vector Layer Generation:

[0173] Based on the regional soil pollution health risk assessment results set, and according to the preset risk level classification thresholds of comprehensive carcinogenic risk value and comprehensive hazard quotient index, the risk level of each dynamic exposure assessment unit is assessed. The risk level is usually divided into three levels: low risk, medium risk, and high risk. Low risk corresponds to a comprehensive carcinogenic risk value lower than the first threshold and a comprehensive hazard quotient index lower than the second threshold. Medium risk corresponds to a comprehensive carcinogenic risk value between the first and third thresholds and a comprehensive hazard quotient index between the second and fourth thresholds. High risk corresponds to a comprehensive carcinogenic risk value higher than the third threshold or a comprehensive hazard quotient index higher than the fourth threshold.

[0174] The assessed risk level attributes are linked to the corresponding spatial elements in the dynamic exposure assessment unit vector layer through spatial association technology, generating a dynamic exposure assessment unit risk zoning vector layer with risk level attributes. During the association process, it is ensured that the risk level attributes and spatial elements correspond one-to-one, without mismatch or omission. This layer simultaneously contains the spatial boundary information and risk level attributes of the dynamic exposure assessment unit, providing core vector data for subsequent visualization rendering.

[0175] Step S5-6: Generation of regional soil pollution health risk level distribution map:

[0176] Symbolic rendering is performed on the vector layer of risk zoning for dynamic exposure assessment units. Differentiated rendering rules are set according to risk levels, such as green for low-risk areas, yellow for medium-risk areas, and red for high-risk areas. Corresponding legend labels are also configured for different risk levels. During the rendering process, the visual presentation of the layers is optimized to ensure that the boundaries of areas with different risk levels are clearly distinguishable, the color transitions are natural, and there is no visual confusion. After rendering, a regional soil pollution health risk level distribution map covering the target area is generated. This map intuitively presents the spatial distribution pattern of soil pollution health risks within the target area, making it easy to quickly identify high-risk cluster areas.

[0177] Step S5-7: Identification and Association of High-Risk Area Dominant POI Categories:

[0178] For dynamic exposure assessment units that are rated as high-risk in the risk zoning vector layer of dynamic exposure assessment units, a spatial overlay analysis tool is launched to overlay their spatial range with the POI spatial database; by statistically analyzing the proportion of the number or density of various POIs in the overlay area, the POI category with the highest proportion is identified and extracted as the dominant POI category of the high-risk unit.

[0179] In the regional soil pollution health risk level distribution map, high-risk dynamic exposure assessment units are marked using a labeling tool. The labeling content includes the dynamic exposure assessment unit identifier, comprehensive carcinogenic risk value, comprehensive hazard quotient index, and dominant POI category information. The labeling location is selected at the geometric center or a prominent location of the high-risk area to ensure that the labeling information is clearly visible and does not obscure key map elements. By associating the labels, the risk level and potential causes of high-risk areas are clarified, providing a clear direction for the accurate formulation of pollution prevention and control measures and risk management strategies.

[0180] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for regional soil pollution health risk assessment based on POI data, characterized in that: Includes the following steps: Step S1: Acquisition and preprocessing of multi-source heterogeneous data: Acquire soil pollution sampling point data, digital elevation model geospatial data, population census grid data, and Points of Interest (POI) data including residential, educational, medical, catering, and park information for the target area; classify, clean, geocode, and spatialize the POI data to establish a POI spatial database; Step S2: Spatial distribution simulation of soil pollutants: Based on soil pollution sampling point data, spatial structure analysis and outlier detection are performed to generate a standardized sampling point dataset that meets the requirements of spatial interpolation. Based on a standardized sampling point dataset, a variogram model is constructed to characterize the spatial autocorrelation features of specific pollutant concentrations. Based on the variogram model, the Kriging spatial interpolation algorithm is used to make the optimal unbiased estimate of the pollutant concentration at unknown spatial locations within the target area, and generate a spatial prediction point set of pollutant concentration. The spatial prediction point set is rasterized to generate a preliminary raster layer of pollutant concentration spatial distribution with continuous concentration values ​​covering the entire target area. Spatial reference and pixel attribute standardization processing are performed on the preliminary raster layer of pollutant concentration spatial distribution to generate a concentration spatial distribution raster layer that is spatially registered with the geospatial data of the digital elevation model and the subsequent surface spatial registration of the population activity intensity index. Step S3: Construction of a dynamically weighted population activity intensity surface: Based on the POI spatial database, POI spatial locations belonging to residential, educational, medical, catering, and park categories are extracted respectively; For each POI category, an initial activity intensity weight value is assigned to represent its relative importance based on its attraction and agglomeration effect on the daily activities of the population in the target area. For each category of POI spatial point set, a kernel density estimation algorithm is used independently for spatial smoothing to generate a density distribution raster layer that reflects the degree of spatial agglomeration of various types of POI facilities in the target area. The search radius of the kernel density estimation is set based on the average service range or spatial distribution characteristics of the POI facilities of that type. The generated density distribution raster layers of various POIs are normalized to ensure that their density values ​​are within a uniform range, thus forming standardized density distribution raster layers of various POIs. The initial weights of various POIs are dynamically corrected based on census grid data. The corrected weights are then weighted and superimposed with the corresponding density layers to generate a population activity intensity index surface that comprehensively reflects the spatiotemporal distribution characteristics of the population. Step S4: Dynamic generation of exposure assessment units and calculation of pollution exposure dose: The population activity intensity index surface is divided into three levels of activity areas: high, medium and low, according to a preset threshold, and the spatial boundary of each level of activity area is used as a dynamic exposure assessment unit; for each exposure assessment unit, a concentration spatial distribution raster layer is superimposed, and the average pollutant concentration within the unit is extracted; combined with preset exposure parameters, the daily average pollutant exposure dose for different age groups within the unit is calculated for the three routes of oral ingestion, skin contact and inhalation. Step S5: Multi-pathway health risk quantification and visualization: Substitute the pollutant exposure doses calculated in each exposure assessment unit and the corresponding pollutant toxicity parameters into the health risk assessment model to calculate the carcinogenic risk value and hazard quotient of each unit; map the calculated risk values ​​back to the corresponding dynamic exposure assessment unit spatial range to generate a regional soil pollution health risk level distribution map, and identify the dominant POI category information in the high-risk area.

2. The method for assessing the health risk of regional soil pollution based on POI data according to claim 1, characterized in that: Step S1 specifically includes: Soil pollution sampling point data is obtained through environmental monitoring networks, digital elevation model geospatial data and population census grid data are obtained through geographic information platforms, and original points of interest (POI) data including residential, educational, medical, catering, and park information are obtained through commercial map service interfaces or public datasets. The original Points of Interest (POI) data is classified and cleaned. Based on the preset POI category system, the attribute information of each POI record is classified and filtered to remove duplicate, erroneous, and irrelevant POI records. The cleaned POI records are geocoded and spatialized, converting the text address information of each POI record into precise geographic coordinates and unifying them to the same spatial reference coordinate system as the digital elevation model geospatial data and the census grid data. Based on the spatialized POI records, geographic coordinates, and corresponding attribute category information, a structured POI spatial database is constructed. This database supports fast querying and statistical analysis based on spatial location and POI category, and provides an input data source for subsequent kernel density estimation algorithms. 3.The method of assessing health risk of regional soil pollution based on POI data according to claim 1, characterized in that: The initial weights of various POIs are dynamically corrected based on census grid data. The corrected weights are then weighted and superimposed with the corresponding density layers to generate a population activity intensity index surface that comprehensively reflects the spatiotemporal distribution characteristics of the population. Specifically, this includes: The standardized POI density distribution raster layers are spatially overlaid with census grid data to calculate the spatial correlation between the normalized density values ​​of various POIs in each grid cell and the number of permanent residents in the grid. Based on the spatial correlation analysis results, a dynamic weight correction model is established. For POI categories that show a significant positive correlation with population size, their initial activity intensity weight values ​​are increased. For POI categories that show no significant correlation or a negative correlation, their initial weight values ​​are maintained or decreased, thus obtaining a set of dynamically corrected POI category activity intensity weight values. The dynamically corrected POI category activity intensity weight values ​​are then multiplied by their corresponding standardized POI density distribution raster layers to generate weighted activity intensity raster layers for each category. Spatially overlay and sum the weighted activity intensity raster layers for each category to generate a surface raster layer that comprehensively reflects the population activity intensity index caused by the combined effects of population distribution and multiple types of POI facilities.

4. The method for regional soil pollution health risk assessment based on POI data according to claim 1, characterized in that: The population activity intensity index surface is divided into high, medium, and low activity zones according to a preset threshold, and the spatial boundaries of each activity zone are used as dynamic exposure assessment units; specifically including: Extract the cell values ​​of the surface raster layer of the population activity intensity index and calculate its statistical quantiles; Based on statistical quantiles, two thresholds are set to divide the population activity areas into high, medium and low levels. The threshold for the high activity area is greater than or equal to the first high quantile, and the threshold for the medium activity area is between the first low quantile and the second low quantile. Based on two thresholds, the surface raster layer of the population activity intensity index is reclassified to generate a population activity intensity partition raster layer with high, medium and low attributes. The raster layer of population activity intensity is converted from raster to vector, and the continuous spatial boundaries of the high, medium and low activity zones are extracted to generate spatial vector boundary files as the basis for subsequent calculations. The spatial vector boundary files are topologically checked and corrected to ensure that the boundaries of each level of activity area are non-overlapping and seamless. The corrected spatial vector boundaries of each level of activity area are defined as high, medium and low dynamic exposure assessment units, forming a dynamic exposure assessment unit vector layer.

5. The method for regional soil pollution health risk assessment based on POI data according to claim 4, characterized in that: For each exposure assessment unit, a concentration spatial distribution raster layer is overlaid to extract the average pollutant concentration within the unit; combined with preset exposure parameters, the average daily pollutant exposure dose for different age groups within the dynamic exposure assessment unit is calculated for oral ingestion, skin contact, and inhalation. Specifically, it includes: Spatial overlay analysis was performed between the dynamic exposure assessment unit vector layer and the concentration spatial distribution raster layer; For each dynamic exposure assessment unit, the arithmetic mean of pollutant concentrations of all raster cells that fall completely within the boundary of the dynamic exposure assessment unit is calculated using a zonal statistical tool. The arithmetic mean of pollutant concentrations is defined as the representative average pollutant concentration of the dynamic exposure assessment unit. Based on the activity area level corresponding to the dynamic exposure assessment unit and the dominant POI category within it, and combined with the typical activity patterns of different age groups in residential, educational, medical, catering, and park POI locations, a set of exposure parameters matching the population activity characteristics of the dynamic exposure assessment unit are determined and assigned. The exposure parameters include at least the daily stay time of different groups in each location, soil ingestion rate, skin exposure surface area and soil adhesion coefficient, and respiration rate. The representative average pollutant concentration of the dynamic exposure assessment unit, the exposure parameters that match the population activity characteristics within the dynamic exposure assessment unit, and the specific calculation formulas for different exposure routes are combined to independently calculate the average daily pollutant exposure dose for children and adults within the dynamic exposure assessment unit via oral ingestion, skin contact, and inhalation. The daily average pollutant exposure doses calculated from three exposure pathways within each dynamic exposure assessment unit and for different population groups are aggregated to form a set of pollutant exposure dose results from multiple pathways and for multiple population groups, with the dynamic exposure assessment unit as the basic recording unit.

6. The method for regional soil pollution health risk assessment based on POI data according to claim 1, characterized in that: The pollutant exposure doses calculated in each exposure assessment unit and the corresponding pollutant toxicity parameters are substituted into the health risk assessment model to calculate the carcinogenic risk value and hazard quotient for each unit. Specifically, it includes: Based on the exposure dose results set of pollutants from multiple routes and multiple populations, the toxicity parameters of a specific pollutant are obtained. The toxicity parameters include at least the carcinogenicity slope factor of oral intake, the risk factor of inhalation unit, and the reference dose for each exposure route. For each dynamic exposure assessment unit in the vector layer of the dynamic exposure assessment unit, the average daily oral exposure dose calculated within the unit for both children and adults is combined with the oral intake carcinogenic slope factor to calculate the carcinogenic risk value of the oral intake route for that unit. For each dynamic exposure assessment unit in the vector layer of the dynamic exposure assessment unit, the average daily exposure dose of respiratory inhalation calculated within the unit for both children and adults is combined with the risk factor of the respiratory inhalation unit to calculate the carcinogenic risk value of the respiratory inhalation route for that unit. For each dynamic exposure assessment unit in the vector layer of the dynamic exposure assessment unit, the average daily exposure doses calculated within the unit for both children and adults for oral ingestion, skin contact, and inhalation are combined with the reference doses for the corresponding exposure routes to calculate the hazard quotient for each route. For each dynamic exposure assessment unit, the carcinogenic risk values ​​of oral ingestion and inhalation are summarized to obtain the comprehensive carcinogenic risk value of the unit; the hazard quotients of oral ingestion, skin contact and inhalation are summarized to obtain the comprehensive hazard quotient index of the unit. Based on the comprehensive carcinogenic risk value and comprehensive hazard quotient of all dynamic exposure assessment units, a set of regional soil pollution health risk assessment results is constructed.

7. The method for regional soil pollution health risk assessment based on POI data according to claim 6, characterized in that: The calculated risk values ​​are mapped back to the corresponding dynamic exposure assessment unit spatial range to generate a regional soil pollution health risk level distribution map, and the dominant POI category information in the high-risk area is associated for identification. Specifically, it includes: Based on the regional soil pollution health risk assessment results set, and according to the preset risk level classification thresholds of comprehensive carcinogenic risk value and comprehensive hazard quotient index, the risk level of each dynamic exposure assessment unit is assessed. The assessed risk level attributes are associated with the corresponding spatial elements in the dynamic exposure assessment unit vector layer to generate a dynamic exposure assessment unit risk zoning vector layer with risk level attributes. Symbolic rendering is performed on the vector map layer of risk zoning of dynamic exposure assessment units, and different colors and legends are assigned according to the risk level to generate a regional soil pollution health risk level distribution map. For dynamic exposure assessment units that are rated as high-risk in the risk zoning vector layer of dynamic exposure assessment units, their spatial range is overlaid with the POI spatial database to identify and extract the dominant POI category within the spatial range of the unit. In the regional soil pollution health risk level distribution map, high-risk dynamic exposure assessment units are marked, and their dominant POI category information is used as auxiliary information for risk causes for association identification.

Images